Lessons from Leaders on Cultivating Culture

Not many people have the opportunity to listen to 8 successful technologists and businesspeople talk about their successes and failures, let alone having the chance to do so in one place at one time.  Below are some notes from CultivateCon 2013, which offered exactly that.  Among our key takeaways:

  • Culture is not a replacement for Product.  The best culture in the world can't make people buy a bad product.  Every speaker touched on this idea to some degree.
  • Transparency is important.  The best people, especially at a startup, are the ones who function best with context and full information.
  • Bored people quit.  Almost everyone in a technical organization has a state change every 3 years or so.  If you're not giving people the opportunity to do that within your organization -- a new position, a new product, a new challenge -- you're risking having your best people walk away.
  • Kittens.  You have to read the notes if you want to understand this one...

Check out the full notes below!

Culture is something that we feel strongly about here at HubSpot.  It's something that we spend a lot of time talking, thinking, and writing about, as well as living on a day-to-day basis, and so we're not unbiased observers.  Take a look at our Culture Code to get a sense of what our lens looks like.

With that background, the speakers:

Tim O'Reilly

Topic: "How I Failed"

Largely similar to a recent blog post, but with a few additional personal anecdotes that highlighted his points.

  • Failure #1: Not making sure that the team actually heard what I was saying.  
    •  That is to say, you can tell people exactly what you want, but you need to take the time to make sure that the thing that they actually hear is the thing that you mean.  
    •  Michael Lewis, talking about the fact that a lot of people went into finance because of Liar's Poker: "you never know what book you've written until you know what book people have read."
    • Have a mission.  Have a set of strategic priorities that further the mission.  Have a set of priorities that contribute to achieving those strategic priorities.  Have a set of activities that are aligned with those priorities.
  • Failure #2: That's how it's done.  
    • Don't be locked in to the way that things have been done historically.  
    • We apply creative thinking (hopefully) to technology problems, do the same for business problems.
  • - Failure #3: Lack of financial and operational discipline. This was probably our biggest failure.  
    • We missed so many opportunities because we didn't have the savvy to understand that we needed to be financially disciplined, we needed to be operationally smart.  
    • O'Reilly almost died as a company because they weren't smart about how they used their cash, they weren't smart about leveraging/negotiating with suppliers.  They looked around after the web bubble burst in 2000/2001 and realized that they had accumulated people who weren't actually doing anything valuable to the priorities->strategy->mission tree.  
    • Treat your financial team as a first-level group.
    • Hold teams accountable for their numbers, make sure that the right team is being measured for any given number
      •  We used to do something silly, where the people putting on a conference weren't the ones who were responsible for the P&L -- that was someone else's problem.  But that meant that one team (the financial team) was being measured based on the success of another (the content team creating the program).  Make one team responsible for both, measure them based on their own success.
    • Run lean, and when you pivot, reassess your staff
      • It's easy to ignore the fact that some people are good at one thing and not at another, and that when you change what it is that you do you might have some people who aren't the right ones.
      • People want Autonomy, Mastery, Purpose (from "Drive" by Daniel Pink).  Keeping someone around who's the wrong one for the job isn't good for you or for them.
  • Failure #4: Don't tolerate mediocrity
    • Sometimes, insisting on taking more time, or spending more money, is the right answer.  If your gut is telling you that something isn't good enough, then it's not good enough.  Go fix it.
  • Failure #5: Hiring supplements, not complements
    • Don't hire more of what you have.  An organization that is monotonically aligned with the person at the top, it's going to be inefficient in some ways, because it's going to be blind to some of what's happening.
    • The person at the top of the tree probably shouldn't be the best at anything the business does
    • O'Reilly had mediocre financial people for a long time, for example
  • Failure #6: "I'll Take Care Of That"
    • When people who work for us aren't strong enough, the instinct is to compensate.
    • This is something that a lot of strong managers, particularly in technical fields, are prone to.  
    • If you do this, you're solving the wrong problem.

Elaine Wherry

Topic: "Cracking Culture"
  • Clarity in communicating your mission, vision, and strategy is key. 
  • Artifacts of culture are not the same as culture.  Good job, you have a ping-pong table.  That does not mean that your company is place where people want to hang out and play ping-pong.
  • A quote that really jumped out: "Meebo is not a lifetime employer, but we want people to grow and become more employable as a result of their time here."  That's quite a statement, one that is hard to swallow at one level, but I think more accurately reflects the state of the employee/employer relationship than most.  The days of the flannel-suit IBM model are pretty much gone.  (This was also echoed, in its own way, by Rands in his talk at the end of the day.  His form of this sentiment is best paraphrased as "people get bored in 3 years.  Bored people quit."
  • "Truth" is vital, but it's not necessarily the same for everywhere.
  • Culture is not a cult -- it's just a way of doing business.  Avoid being self-congratulatory.  (Yes, HubSpot sometimes gets self-congratulatory about our culture.  This was a valuable takeaway.  We need to be sure that we're constantly re-examining.)

Kate Matsudaira

Topic: "What _DO_ you do all day?"
  • How do you measure performance?
    • Hours worked?  Means nothing
    • Lines of code?  Doesn't account for how difficult or meaningful that code is
    • Bugs?  Tests?  Don't account for customer satisfaction
    • Features?  Doesn't account for quality
  • I've been thinking a lot about flat organizations.  An organization with no managers is an organization where everyone is a manager, because someone still needs to lead.  
  • Leadership is a decision, not a position.
  • 3 types of informal power:
    • Charisma: Charisma is a thing that you have or you don't.  It's not something that can be taught, really (though it is something that you can optimize)
    • Expertise: Lots of people end up in a leadership role because they're simply the most knowledgable about something.  This, too, isn't really something that can be taught.
    • Relationships
      • Start from trust
      • What is trust in the context of an organization?  
      • Don't expect things from others that you wouldn't ask of yourself
      • Know your timing -- whether in the context of a day, or a deliverable
  • Relationship Architecture
    • Make two lists: The most important people on your team; and, the people on the team with whom you have the best relationships.  How much overlap is there?
      • How can you strengthen those relationships?
      • If the people on the second list aren't the same as the ones on the first, why not?  Why aren't you making a point to build relationships with the people who are leaders in their own right?
    • Rebuild burned bridges.  Don't bypass a problem, address it.  
    • Don't deflect, especially in the context of conflict.  Sincerity in your responses to conflict will build other people's trust in you.
    • Be present.  Stay focused.
    • Ask questions, wait for answers.
    • Focus on the long term.  Who are the people who you should have relationships with?  What can you do to develop those relationships, and to nurture them?
  • Things won't always go well
    • It's easy to forget how hard relationships are.
  • Where does success come from?  From people.  Ideas are cool, things that are actually 

Patrick Collison

Topic: "Nothing to Hide: Living with complete email transparency"
  • Stripe has rethought everything from scratch, haven't adopted much.  This has ended up being a little inefficient.  For example, we initially thought that we didn't need an office, turned out we did.  Initially thought that everyone should get paid the same, that turns out to be silly.
  • Rethinking everything from first principles can be exhausting and inefficient, but sometimes you stumble on something that works.
  • I'm a big believer in ambient transparency.  The transparency that comes of having a small org in a small space, but that you lose as you grow.
    • Initially, when the company was only a couple of people, just used a single shared email address.  As we hired new people, we added them as well (though gave them their own mailboxes)
    • Set up outgoing mail server to automatically BCC everyone else in the company on outgoing email
  • That doesn't scale forever, of course.  But, we felt that it was important that everything was internally public.  
    • We created mailing lists for every functional group (sys, ops, dev, investors, etc, etc), and BCC outgoing mail there.
    • Created tooling for automated configuration of GMail filters to avoid overloading people
  • Benefits
    • Means that people are not surprised by things happening in the company.  Surprises are generally not good for morale
    • Makes it easier to integrate new folks
    • Less communication needs to happen, sometimes.  Less time spent talking to X or Y trying to find out the state of a project or something like that.
    • Eases people's curiosity
    • Helps in reducing politics.  
  • We have the goal of being a "hive mind", and we push our tools that way.  For example, we use Hack Pad -- a shared text editor that will even give you a feed of documents being used across the organization.  It's another version of the idea of overhearing a conversation in a 4 person office.
  • Decide what properties you want in your communication.  Diff that against the way that you communicate now.  Solve for the diff.
  • Some other lessons: 
    • You should in fact have an office
    • Product Management may not be a useful function -- have engineers manage the product
    • In hiring, self-discipline is maybe the most important thing
    • Hiring someone who is capable of being very good at doing their thing well is undervalued.  If hiring engineers, code with them all the time, and see how well they do with that.
    • Hack Trips -- changing scenery, taking some people who code to somewhere else can be a great way to help productivity.  (Stripe has done these to places like Rio and Hawaii, it's something that they brought over from Github.)
    • Have a shipped@ email address, and every time something gets shipped, send an email that announces it.  Focus on quantifying the impact, not just celebrating the fact that a thing was done.
    • When you fire someone, be very open about the reasons why, and be honest about them.  People will know if you are not telling the truth.
    • Thank people when they raise a concern about the company.
    • have lunch and dinner at the office.  Hire a chef.  (It makes financial sense around 20-25 people.)
    • http://bit.do/cultivate
  • Audience Question: Does transparency like this cause problems given that Stripe is handing financial data and PII?
    • We keep user data and company data very separate.  User data is tightly controlled, and the expectations are very clear about it.  Company data is very open, though.

Hiten Shah

Topic: "How to Create a Culture of Shipping Code Continuously"
  • Hiten has a reputation for dropping "founder bombs." That is, he gets out of a conversation, meeting, or reading session, has some great ideas, and calls his team to share them.
  • He had to learn to make sure the people hearing these ideas would take them with a grain of salt and not just as commandments.
  • Teams should know what they're working on, and why. The "why" shouldn't be "because my boss told me to"; it should be something like "to reduce churn" or "to affect these metrics."
  • Closed door decisions lead to people doing things "because my boss told me to". Need to pass down understanding of the problem, the goal, and how the decision was made.
  • Build products people want. -Paul Graham
  • Seek feedback from your team.

Scott Chacon

Topic: "Leading by First Priciples".  (Side note: Scott Chacon looks a lot like an old roommate of mine.  It's kind of distracting at first.)
  • An anecdote.  100 years ago, a yarn factory had poor morale and low productivity.  An organizational psychologist suggested kittens -- have kittens running around playing with the yarn.  It worked, people were happier and productivity went up.  The most amazing part is that some factory owner said yes to this idea.  The morale: if someone suggested "hey, let's try kittens", would you say yes?
  • First principles -- things that are more important than other things.  Disagreements should always come back to these.  Once you agree on principles, the rest follows from that.
    • "Three-year old prosecution": 3-year olds just want to know why.  "I think we should do this".  Why?  "Because doing X means Y, and Y is important."  Why?...
    • This is also the 5-whys approach.
  • What are the principles of your company?  Unlikely to be the same as Github's "open source business" model
    • The office.  Why do we need an office?  The owners of the cafe where we're working are getting mad.  First Principles:
      • Needed a place to work
      • Needs to foster creative collaboration
      • Needs to encourage serendipitous interactions
      • Needs to allow for "real business" interactions
      • Need a place to physically store and ship things
    • The schedule.  Why do we need a schedule?  We don't -- with employees all over the world no one schedule is going to make sense for everyone.  First principles:
      • People need to get their job done, and that's really it.  When they get their job done isn't particularly important.
    • Meetings.  Why do we need to have meetings?  We don't -- for the same reasons as the schedule, meetings don't really work.  If you really need to have a meeting, it needs to be done in some form that is distributed and asynchronous.  First principles:
      • Make it easy for people to get their job done.
      • At a distributed company, you have to go out of your way to make sure that you are including the people who are remote in time or space.
    • Management: Why do we need management?  We don't.  The overlap of "management" and "problems that Github has" is very small.  We have chosen to distribute these functions instead of hiring professionals for the moment (coordination, planning, strategy, etc).  Give people trust, that they will make intelligent decisions about where to spend their time.  First principles:
      • Give trust.
      • Don't solve for problems that you don't have.
  • Don't cargo cult.  If you're going to do something, there had better be a reason that you are doing something.  Understand that reason.

Patty McCord

Topic: "Leveraging Logic As a Leader"
  • Fundamentals:
    • Company First.  You don't need a great culture if your product sucks or your idea sucks.  Culture enables success, but it does not cause success.
    • Judgment trumps everything.  Fairness is important, consistency is not.  Use your judgment of what's fair, use your judgment of what's smart, use your judgment of what's right.  Don't worry about being consistent with all of the things that you've thought in the past.
    • Just tell people the truth.  
      • What does a strong team look like?
        • You can't really answer this without knowing what you're building.  What does that team need to do?  The answer to that determines who you actually need.
        • If the people you have can't get the job done, tell them that.  It's not being mean, it's being honest.
        • It gives people the chance to decide if they're motivated by the challenge of the work, it gives them the choice of deciding what they want their own career to look like.
      • My least favorite type of person is the whiner.  "Oh, Patty, things have changed, the culture is different and I don't know if management realizes it."  And I look at them and I say "Yep, I'm management, and I'm very aware of it."  And they say "Why?  Why does it need to be this way?"  And I look back at them and I say "Because we're successful.  And we want to be a big, successful corporation.  And that means that things need to be different than they were."
      • Imagine a world with no Performance Improvement Plans.
        • Just tell the truth!
      • Imagine a world where interviewing is not a sin
        • When you're talking to someone who isn't happy or isn't performing, suggest that they go interview
          • Turns out the grass isn't always greener
          • When they decide to stay, then you can ask what they offered.  That tells you what actual market comp is.
    • Visualize success
      • In 6 months, if you had the perfect team in place, what would be happening that isn't now?
      • What would you measure?
        • Metrics?
        • Deadlines?
        • Progress?
      • What would it look like?
      • What would people have to know to pull this off?
  • So quotable!
    • "Someone was saying that they moved everyone's chairs in the office, and they all freaked out.  Classic geek!  So sometimes I would move their chairs just to fuck with them."
    • "I was jealous because everyone else at Netflix got to innovate, and I didn't.  So I decided I'd start.  And the way that I innovated was by just not doing shit."

Michael Lopp (aka Rands in Repose)

Topic: "Why I Hate Meetings"
  • You are (at most) three years from building something new.  Look at resumes in tech.  Everyone has a state change every 3 years (or so often as to be effectively all the time).  
    • Bored people quit
    • Get a report, every Monday -- everyone who is at 2 years.  What they're working on.
  • Irrelevancy is just around the corner.  There is someone out there who is making everything that you do unimportant, or commodity.  It's going to happen, and it's good.
  • I hate meetings.
    • Why did meetings come into existence (a theory)
      • Scaling communication
      • At some point, we (as leaders) no longer know when things are going off the rails.
      • A timeline:
        • One person in a cafe has an idea and does some analysis and decides it makes sense, so he starts to work on it
        • Hires someone else, describes the idea, they debate it and make it better and then go work on it.  And sometimes they disagree, and they talk about it, and then they come to an agreement.
          • This defines a meeting
            • Not just you
            • Together working to solve a problem
            • In a finite amount of time
            • As needed
          • But sometimes this is not what happens
        • They hire a bunch more people, and they build the company, and it's up to 30 people.  Ideas are coming from every direction.  Decisions are collaborative and visible.  Execution is key, and there's no time to bicker.  Everyone knows everyone and does everything.  Error correction is rapid and organic.  Lost cost situation awareness.  The most random stuff becomes culture.  (This is amazing, and is also the establishment of "the old guard")
          • The idea of a meeting is still pretty consistent.
        • Now the company grows some more.  Now it's 300 people.  Ideas still show up but there's an overwhelming number.  Decisions are slower.  Execution is stove-piped.  (This is "the new guard", and they're pissed, because the old guard can flit around and just do things and it seems so damn painless.)
          • Stories from the past become myth
          • Situational awareness is expensive
          • Communication requires overcoming friction
          • Learning can no longer occur via osmosis
        • Things go off the rails
          • The old guard has no incentive to actually make things systematic, because they are the special few who can fix things by wandering the halls
          • But someday something blows up (Palantir example: "Steve" quit -- the best engineer, the guy who does all the things.  And the old guard didn't know, because of a signal/noise ratio issue.)
          • A meeting _can_ actually solve things, sometimes, and in this case it does
        • Meetings go viral
          • Perceived value of meetings go up
          • It begins to be seen as the right way to solve things
          • People who are really good at running a meeting are incentivized to continue having meetings
        • Meetings turn into something else
          • Random people who aren't contributing
          • A lack of agenda
          • That doesn't respect the attendee's time
          • That goes on forever and ever and ever and ever
  • How do we prevent meeting culture?
    • Have more meetings!  Obviously.
    • Actually, have a culture that values consistent, useful communication in all directions
  • Preventative Maintenance Meetings
    • 1-on-1's
      • Every week
      • No matter what
      • 30 minutes (at least)
      • Cheat sheet
        • Performance review: "You said you wanted to be able to do X, how's that going?"
        • A current disaster: "Here's something blowing up in my world, how would you handle this?"
    • Staff meeting (1-to-many): At Apple, it's called "E-Team", and if you get invited your thing is fucked up.  The point: Qualitative information is important too. 
      • Every week (early)
      • No matter what
      • 30 minutes (at least)
      • Cheat sheet
        • The Dashboard: What is the health of the team?  What is the health of the business?
        • Disasters: Things that are blowing up right now
        • Special Guests: Someone else from around the company comes in to share the dashboard and disasters from their team.
    • Tapestry Meeting (many-to-many)
      • Occasional
      • Possibly over poker
      • Not only when things are on fire
      • The point: To set up an environment where unexpected/serendipitous conversations happen
  • All other meetings need an expiration date

Ian Marlier

Written by Ian Marlier

Director of Reliability, HubSpot

Comments

Subscribe for updates

New Call-to-action